Erich Liang

Department: Computer Science
Faculty Adviser: Jia Deng
Year of Study: G2
Undergraduate School: Caltech
Undergraduate Major: Computer Science and Math

Personal Bio

Hi! I'm Erich Liang, a current second year CS PhD student studying 3D computer vision. I've been a west coast person for most of my life (grew up in Seattle area and did my undergraduate studies at Caltech), so this will be my first time living on the east coast for an extended period of time. As an undergrad, I was unsure about what I wanted to do at Caltech and after graduation, so if this sounds like you, there's nothing to worry about! Personally, I found exploring different career paths first hand helped immensely. If you'd like to explore the possibility of doing research / academia with me, I'd be happy to show you the ropes :) Looking forward to meeting you!

(If you'd like to chat and we don't get a chance to meet in-person, feel free to reach out to me at erliang@princeton.edu)

Fun Fact

My name ends with an "h" at the end. It is supposed to be silent, but if you choose to pronounce it, you can come up with many interesting nicknames for me...

Research Pitch

Key terms: 3D computer vision, computer graphics, human vision

Research pitch: As humans, we are able to perceive the world around us in full 3D, but in reality, our eyes only provide 2D visual input of our surroundings. Our brains are able to take in these 2D signals and "stitch" them into our 3D perception; can we design computer algorithms to do the same? First, I am interested in the problem of robust 3D reconstruction, with particular emphasis on recovering accurate geometry and material properties from input 2D images of an object. Humans are also able to guess 3D properties of an object without having seen it from every possible viewpoint, and I am interested in helping computer algorithms learn "3D intuitions" about the world around us from 3D data at scale.

If any of this sounds interesting, feel free to reach out (erliang@princeton.edu)! The field moves quite fast, but current trending methodologies include Neural Radiance Fields (NeRFs) and Gaussian Splatting. Research in this area typically involves coding in Python and/or C++, with some occasional image dataset collection by hand.

Upcoming Programs That I Am Attending:

Plans for Summer 2025

Interested in participating in Summer 2025 ReMatch+ program.